The computation of the maximun loading point (MLP) is crucial to power systems operation and control, There are several methodologies proposed to compute it. The continuation load flow (CLF) is very robust, widely known to draw PV curves, and can be also used for computing the MLP. However it has some drawbacks, the procedure may diverge for some cases and it is very conservative for some networks, taking many iterations. Other efficient technique using nonlinear programming (NLP) has been proposed recently. This method presents clear advantages due to the orientation of the process in direction to the MLP. This paper presents a method based on the CLF and NLP . The idea is to compute the MLP taking some features of the CLF combined with characteristics of NLP. With this combination the MLP can be evaluated with more accuracy and efficiency. Simulations for different systems including IEEE test systems are shown to evaluate the performance of the method. Some comparisons of the methodologies are also shown. ©2007 IEEE.